Information Systems Students' Study Performance Prediction Using Data Mining Approach

Authors:
Publisher website
Google Scholar
Abstract
The quality of study program performance can be seen from the achievement of the accreditation status of the study program, where one element of the assessment instrument related to graduate profile is the study performance. Graduate on time is one of the indicators of a student's success in obtaining his undergraduate degree and becomes an important attribute, because by being able to predict the study period of students, higher education institutions can minimize the failure of students graduation by making planning, escorting studies, and more intensive guidance. If the data stored in the academic database is only processed for reporting purposes and has not been utilized optimally to support decision making, it is very unfortunate because the so large data are not used to extract what information are contained in the data. Regarding students on time graduation prediction, data mining classification can be performed against the data stored in academic database. This research aims to predict the students on time graduation using data mining decision tree classification with C4.5 algorithm and to find out what attributes influence the prediction. The attributes used consisted of gender, hometown, school hometown, shift, entrance exam grade, and Grade Point Average (GPA). The result showed that the most influential attribute was the GPA and C4.5 algorithm had a performance in classifying the data with an accuracy value of 78.612%, a precision value of 0.762, and a recall value of 0.786.

This publication has 25 references indexed in Scilit: